English

Transformer Approximations from ReLUs

Machine Learning 2026-04-29 v1 Artificial Intelligence Machine Learning

Abstract

We provide a systematic recipe for translating ReLU approximation results to softmax attention mechanism. This recipe covers many common approximation targets. Importantly, it yields target-specific, economic resource bounds beyond universal approximation statements. We showcase the recipe on multiplication, reciprocal computation, and min/max primitives. These results provide new analytical tools for analyzing softmax transformer models.

Cite

@article{arxiv.2604.24878,
  title  = {Transformer Approximations from ReLUs},
  author = {Jerry Yao-Chieh Hu and Mingcheng Lu and Yi-Chen Lee and Han Liu},
  journal= {arXiv preprint arXiv:2604.24878},
  year   = {2026}
}
R2 v1 2026-07-01T12:37:57.436Z